X2Paddle
X2Paddle copied to clipboard
PaddlePaddle
Message type "caffe.DropoutParameter" has no field named "scale_train"
感谢您参与 X2Paddle 社区！ 问题模版为了 X2Paddle 能更好的迭代，例如新功能发布、 RoadMaps 和错误跟踪. :smile_cat:
问题描述
- google.protobuf.text_format.ParseError: 318:5 : Message type "caffe.DropoutParameter" has no field named "scale_train".
- INFO:root:paddle.version = 2.1.0
INFO:root:Now translating model from caffe to paddle.
Traceback (most recent call last):
File "d:\program files\python39\lib\runpy.py", line 197, in _run_module_as_main
return _run_code(code, main_globals, None,
File "d:\program files\python39\lib\runpy.py", line 87, in run_code
exec(code, run_globals)
File "D:\Program Files\Python39\Scripts\x2paddle.exe_main.py", line 7, in
File "d:\program files\python39\lib\site-packages\x2paddle\convert.py", line 339, in main caffe2paddle( File "d:\program files\python39\lib\site-packages\x2paddle\convert.py", line 183, in caffe2paddle model = CaffeDecoder(proto_file, weight_file, caffe_proto) File "d:\program files\python39\lib\site-packages\x2paddle\decoder\caffe_decoder.py", line 290, in init text_format.Merge(proto_str, self.net) File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 731, in Merge return MergeLines( File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 805, in MergeLines return parser.MergeLines(lines, message) File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 830, in MergeLines self._ParseOrMerge(lines, message) File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 852, in _ParseOrMerge self._MergeField(tokenizer, message) File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 982, in _MergeField merger(tokenizer, message, field) File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 1057, in _MergeMessageField self._MergeField(tokenizer, sub_message) File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 982, in _MergeField merger(tokenizer, message, field) File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 1057, in _MergeMessageField self._MergeField(tokenizer, sub_message) File "d:\program files\python39\lib\site-packages\google\protobuf\text_format.py", line 947, in _MergeField raise tokenizer.ParseErrorPreviousToken( google.protobuf.text_format.ParseError: 318:5 : Message type "caffe.DropoutParameter" has no field named "scale_train".
具体信息
- 转换模型后用处
- 使用 Paddle 框架/ PaddleInference 推理预测
- 模型来源 caffe模型。
- 应用场景 用于服务端业务下的检测业务
- 版本信息 PaddlePaddle => :grey_question:2.1 X2Paddle => :grey_question:1.3.5 来源框架版本(Caffe) => :grey_question:不清楚
- 您的联系方式(邮箱/微信/电话) 微信：hurui79
caffe模型地址： 链接：https://pan.baidu.com/s/1lKyvx6tplZD7OSXqSomImQ 提取码：zwqm
@390029659 您好，看起来像caffe版本问题，检查一下模型对应caffe版本吧~
可是安装依赖并不需要安装caffe，我需要在本地也安装相应的caffe么？ @wjj19950828
@390029659 您好，看起来像caffe版本问题，检查一下模型对应caffe版本吧~ 您好这回信息又变了（我把protxt 里面的scale_train都注释掉了），求帮忙给看看@wjj19950828 INFO:root:paddle.version = 2.1.0 INFO:root:Now translating model from caffe to paddle. The filter layer:drop6 The filter layer:drop7 Traceback (most recent call last): File "d:\program files\python39\lib\runpy.py", line 197, in _run_module_as_main return _run_code(code, main_globals, None, File "d:\program files\python39\lib\runpy.py", line 87, in run_code exec(code, run_globals) File "D:\Program Files\Python39\Scripts\x2paddle.exe_main.py", line 7, in
File "d:\program files\python39\lib\site-packages\x2paddle\convert.py", line 339, in main caffe2paddle( File "d:\program files\python39\lib\site-packages\x2paddle\convert.py", line 183, in caffe2paddle model = CaffeDecoder(proto_file, weight_file, caffe_proto) File "d:\program files\python39\lib\site-packages\x2paddle\decoder\caffe_decoder.py", line 296, in init self.caffe_graph.build() File "d:\program files\python39\lib\site-packages\x2paddle\decoder\caffe_decoder.py", line 252, in build self.set_node_shape(node) File "d:\program files\python39\lib\site-packages\x2paddle\decoder\caffe_decoder.py", line 277, in set_node_shape node.out_shapes = getattr(caffe_shape_inference, func_name)(node.layer, AttributeError: module 'x2paddle.decoder.caffe_shape_inference' has no attribute 'shape_python'
错误信息这回到我最开始的的时候出现的了
Message type "caffe.LayerParameter" has no field named "input_param" @wjj19950828 就很绝望。。。。
@390029659 您好～
第一个问题，"scale_train"问题将deploy.prototxt中的相关信息注释到即可，因为dropout不影响预测
第二个问题，因为您提供的模型有自定义层proposal，所以必须使用自己环境的caffe_proto
第三个问题，关于“input_param”，这个也是与caffe版本相关，必须使用您使用的fast-rcnn repo,使用相应caffe版本，参考https://github.com/rbgirshick/fast-rcnn/issues/1
但是个人推荐因为模型较为经典：FasterRcnn,因为要使用Paddle部署，可使用PaddleDetection或者PaddleX，训练部署一体化～
@wjj19950828